A novel attribute weighting method with genetic algorithm for document classification

Abstract

Thanks to the proliferation of Internet, a lot of data are produced by both Web sites and personal users. The documents are required to be classified in terms of their content in order to reach the necessary information fast and correctly from produced data. One of the biggest difficulties in document classification systems is detection of attribute that represent the classes in best way. In this research, a new attribute method is presented by using a Genetic Algorithm for document classification problem. This proposed method is tested on 450 documents that are from 6 different categories collected from a news portal that broadcasts online. According to experimental results 93% of success is achieved with the proposed method.